The hope of launching neural networks on wearable devices appeared after the invention of AI-microprocessors by Massachusetts scientists.

A new processor with a three to seven times higher performance has 95% less power consumption than existing analogues. It is these characteristics that allow the installation of energy-consuming neural networks for handheld devices.

According to IBM President Dario Gil, the development of scientists at the Massachusetts Institute of Technology is "a new word in the implementation of ultra-precise operations with huge arrays of memory." The invention opens the way to the creation of more complex and accurate neural networks for portable gadgets of the Internet of things.